Comparative Study of the Immunoregulatory Capacity of In Vitro Generated Tolerogenic Dendritic Cells, Suppressor Macrophages, and Myeloid-Derived Suppressor Cells

Abstract

Background 

Regulatory myeloid cell (RMC) therapy is a promising strategy for the treatment of immunological disorders such as autoimmune disease and allograft transplant rejection. Various RMC subsets can be derived from total bone marrow using different protocols, but their phenotypes often overlap, raising questions about whether they are truly distinct.

Methods 

In this study, we directly compared the phenotype and function of 3 types of RMCs, tolerogenic dendritic cells, suppressor macrophages, and myeloid-derived suppressor cells, generated in vitro from the same mouse strain in a single laboratory.

Results 

We show that the 3 RMC subsets tested in this study share some phenotypic markers, suppress T cell proliferation in vitro and were all able to prolong allograft survival in a model of skin transplantation. However, our results highlight distinct mechanisms of action that are specific to each cell population.

Conclusions 

This study shows for the first time a side-by-side comparison of 3 types of RMCs using the same phenotypic and functional assays, thus providing a robust analysis of their similarities and differences.

In Brief

Carretero-Iglesia et al directly compare tolerogenic dendritic cells, suppressor macrophages and myeloid-derived suppressor cells to show that they can be distinguished by their morphology, phenotype, intrinsic properties and in vitro mechanisms of action on T cells.

Current immunosuppressive treatments after organ transplantation are effective in preventing allograft rejection in the short-term, but long-term graft failure remains an unresolved problem.¹ Moreover, such treatments can lead to immunosuppression-related side effects. In this context, cell therapy arises as a promising strategy to reduce the immunosuppressive regimen and, ultimately, to induce permanent donor-specific tolerance.²

Regulatory myeloid cells (RMCs), such as tolerogenic dendritic cells (DCs) (TolDCs), regulatory/suppressor macrophages (suppMφ) and myeloid-derived suppressor cells (MDSCs), are gaining interest as therapeutic agents due to their ability to modulate effector T cell activity by directly targeting activated T cells or by inducing regulatory T (Treg) cells.^3,4 In animal models, we and others have demonstrated the capacity of rodent TolDCs, MDSCs, and regulatory/suppMφ to prolong graft survival^5-11 and to suppress autoimmune responses.^12-15

Significant diversity has been reported between in vitro derived-TolDCs generated in different laboratories. A seminal study from Lutz and colleagues showed that DCs generated from bone marrow (BM) cells with low doses of granulocyte-macrophage colony-stimulating factor (GM-CSF) display tolerogenic properties in vitro and prolong the survival of cardiac allografts.⁵ Other DC manipulations have been described to increase their efficacy and block the maturation process, such as combining GM-CSF + IL-4 with vitamin D3,¹⁶ IL-10,¹⁷ rapamycin,¹⁸ or dexamethasone/lipopolysaccharide (LPS).^19-21 Donor-derived TolDCs and donor-antigen pulsed recipient TolDCs have been reported to prolong graft survival in various transplantation models. We have previously shown that injection of autologous TolDC (called ATDCs) generated with low doses of GM-CSF prolonged skin graft survival in mice.¹⁰

Myeloid-derived suppressive cells were initially described in pathological situations, such as tumor development or inflammation and represent a heterogeneous population of myeloid cells displaying T cell–suppressive activity.^22,23 In vitro differentiated MDSCs were first described by Rossner et al²⁴ as DC myeloid precursors, derived from BM cells cultured with GM-CSF. Other factors, such as IL-13²⁵ or tumor cell-conditioned medium²⁶ were added to MDSC differentiation protocols to promote their in vitro expansion. Subsequently, IL-6 appeared as a potent complement cytokine to GM-CSF because BM-derived MDSCs generated with GM-CSF and IL-6 display stronger suppressive activity and prolong islet allograft survival.²⁷ We recently confirmed the therapeutic potential of autologous GM-CSF/IL-6 MDSCs in a skin allograft model.²⁸

Lastly, regulatory macrophages (Mreg), described by Riquelme et al,¹¹ have been reported to inhibit T cell proliferation in vitro and prolong allograft survival in a cardiac transplantation model. These Mreg were generated from mouse CD11b⁺ Ly6C⁺ BM cells cultured with macrophage-CSF (M-CSF), human serum, and IFN-γ.

Promising results have been obtained in the first clinical trials evaluating the safety of RMC therapy in transplantation²⁹ and autoimmunity.³⁰ As part of the ONE Study project (http://www.onestudy.org/), we and other groups are currently evaluating the safety and efficacy (phase I/II clinical trials) of different types of regulatory cells in kidney transplant patients.^3,31 Important advances have been made in the classification of myeloid cell types by focusing on the expression of transcription factors and surface molecules.^32-34 However, because RMCs display similar phenotypes and immunosuppressive functions, it remains controversial whether they represent distinct cellular populations.

Here, using well-established protocols, we generated and characterized 3 RMC populations with potential for cell therapy: autologous TolDCs (named ATDCs as in our previous studies^9,10), suppMφ, and MDSCs. We compared their in vitro characteristics under well-controlled conditions, as well as their in vivo potential using a murine model of skin transplantation. We demonstrate that although these RMC populations share some phenotypic and functional properties, the mechanisms underlying their tolerogenic profiles are distinct.

MATERIALS AND METHODS

Mice

C57BL/6 and Balb/c mice (6-8 weeks old) were purchased from Janvier (France). This study was carried out in strict accordance with the protocol approved by the Committee on the Ethics of Animal Experiments of Pays de la Loire (CEEA.2013.9).

Cell Preparations

All RMC populations were derived from BM cells from C57BL/6 female mice. Autologous TolDCs and MDSCs were generated as previously described.^28,35 To generate ATDC, BM cells were cultured at a density of 0.5 × 10⁶ cells/mL for 8 days in 100 mm untreated Petri dishes in 10 ml of complete RPMI 1640 medium supplemented with low doses of GM-CSF (0.4 ng/mL) (from COS supernatant). Ten milliliters of complete medium supplemented with 0.4 ng/mL of GM-CSF were added at day 3 of culture and 10 mL of medium were replaced on day 6. Adherent cells were harvested on day 8. To obtain MDSC, BM cells were cultured for 4 days at 2.5 × 10⁶ cells/mL into 100-mm culture-treated Petri dishes in 10 mL of complete Dulbecco's Modified Eagle medium supplemented with 40 ng/mL of GM-CSF (Peprotech) and 40 ng/mL of IL-6 (Sigma-Aldrich). Suppressor macrophages were generated by culturing 10⁶ cells/mL for 15 days in 100-mm dishes in 10 mL of complete DMEM medium, supplemented with low dose M-CSF (0.2 ng/mL, Peprotech). Ten milliliters of medium supplemented with M-CSF were added at day 3, and 10 mL of medium were replaced on day 7. Control myeloid cells were obtained from BM cells cultured for 8 days in complete RPMI1640 medium with 40 ng/mL GM-CSF and matured by addition of 1 μg/mL LPS (Sigma) during the last 24 hours of culture.

Complete DMEM and RPMI1640 media contained 100 U/mL penicillin-streptomycin, 10 mM HEPES, 1 mM sodium pyruvate, nonessential amino acids (all from Gibco), 2 mM glutamine, and 50 μM β-mercaptoethanol (both from Sigma-Aldrich) and 10% of heat-inactivated fetal bovine serum (from Lonza). Harvesting of adherent RMCs was performed by flushes with cold phosphate-buffered saline, 2% fetal calf serum, and 0.1 mM EthyleneDiamineTetraAcetic acid.

When indicated, 0.5 × 10⁶ cells were stimulated with 1 μg/mL LPS (Sigma-Aldrich) or left untreated for 48 hours in 24-well plates (final volume of 500 μL/well—complete RPMI1640 medium).

Flow Cytometry

Regulatory myeloid cell and T cell staining was performed using monoclonal antibodies from BD Biosciences. Anti-MHC (Major Histocompatibility Complex) class II, anti-F4/80, and anti-CD169 were from eBioscience, and anti-CD205 was from Serotec. Foxp3 staining was performed using Foxp3/Transcription Factor Staining Buffer Set (eBioscience) according to the manufacturer's protocol. Dead cells were excluded by DAPI positive staining or Viability Dye eFluor 506 staining (eBioscience). Male antigen UTY-specific CD8⁺ T cells were detected using a PE labelled Pro5 MHC Pentamer (H-2D_b, WMHHNMDLI) (ProImmune). Data acquisition was performed on a FACS Canto II (BD Biosciences) flow cytometer, and the data were analyzed using FlowJo software (Tree Star, Inc).

Antigen Internalization and Degradation Assays

Endocytosis and degradation were determined using Alexa Fluor 647-conjugated OVA (OVA-AF647) and DQ-OVA (Invitrogen). One million RMCs were incubated in 96-well plates (final volume of 200 μL/well—complete RPMI1640 medium) with OVA protein (1 μg/mL) at 37°C/5% CO₂ and analyzed at different time points as described in the Results section. Cells were washed, fixed with 2% paraformaldehyde and analyzed by flow cytometry. A control at 4°C was included at each time point.

In Vitro Allogeneic Cocultures

Splenic Balb/c T cells were purified (Pan T Cell Isolation Kit II, Miltenyi Biotec), carboxyfluorescein diacetate succinimidyl ester (CFSE)-labeled (Invitrogen) and 10⁵ T cells were cultured with decreasing numbers of C57BL/6 RMCs in 96-well plates (final volume of 200 μL/well—complete RPMI1640 medium). T cell proliferation was measured by CFSE dilution using flow cytometry after 4 days of culture and IFN-γ secretion was measured by enzyme-linked immunosorbent assay (BD OptEIA).

In Vitro Suppression Assay

Carboxyfluorescein diacetate succinimidyl ester–labeled splenic C57BL/6 T cells (10⁵ cells) were stimulated with 4 × 10⁴ Dynabeads Mouse T-Activator CD3/28 (Invitrogen) in the presence of RMCs in 96-well plates (final volume of 200 μL/well—complete RPMI1640 medium). T cell proliferation was analyzed on day 4 by CFSE dilution and IFN-γ secretion was measured by ELISA.

T Cell Restimulation Assay

One million CFSE-labeled Balb/c T cells were cultured with C57BL/6 derived RMC for 3 days (1:1 ratio) in 24-well plates (final volume of 1 ml/well—complete RPMI1640 medium). Balb/c T cells were then isolated again, using mouse CD90.2 magnetic beads (Miltenyi Biotec) and 0.2 × 10⁶ T cells were stimulated with purified C57BL/6 splenic CD11c⁺ DCs (Miltenyi Biotec) in 96-well plates (final volume of 200 μl/well—complete RPMI1640 medium) at a 10:1 ratio. Proliferation was measured by CFSE dilution on day 3, and IFN-γ secretion was measured by ELISA.

Skin Transplantation and Treatments

Male C57BL/6 tail skin grafts were performed and monitored as previously described.^28,36 Recipients were injected either with 10⁶ or 3 × 10⁶ autologous nonpulsed RMCs intravenously the day before transplantation or left untreated.

Statistical Analysis

Statistical analyses were performed using the nonparametric Mann-Whitney U test. Survival curves were compared using the Log-rank (Mantel-Cox) test. Analyses were performed using GraphPad Prism.

RESULTS

In Vitro Generation and Characterization of RMCs

Regulatory myeloid cells were differentiated in vitro from total BM cells of naive mice under different culture conditions. Based on protocols previously described by our laboratory^10,28 and others,^5,27 ATDCs were obtained after culture of BM cells in the presence of low-dose GM-CSF (0.4 ng/ml), whereas MDSCs were derived from BM cells cultured with 40 ng/mL GM-CSF and 40 ng/mL IL-6. To compare cells derived from the same source (total BM) and without prior activation (LPS or IFN-γ stimulation), suppMφ were also derived from total BM cells and cultured in the presence of low dose M-CSF (0.2 ng/mL) as previously described³⁷ (Figure 1A). This protocol of suppMφ generation is slightly different from that described by Riquelme et al¹¹ in which Mreg were derived from monocytes isolated from BM and cultured with M-CSF, human serum, and IFNγ. Control myeloid cells, used as nontolerogenic cells, were differentiated from BM cells cultured with 40 ng/mL of GM-CSF and matured with 1 μg/mL of LPS during the last 24 hours of differentiation.

FIGURE 1: Morphology and phenotype of RMCs. A, Summary of protocols used to generate the three RMCs and control cells from bone marrow cells. B, Morphology of RMCs in culture and HES. C, Percentage of cell recovery for each RMC culture (n = 3). D, Expression of surface markers defining each RMC type. Histograms are representative of 5 independent experiments. HES, hematoxylin-eosin staining.

At the end of the culture, each RMC type displayed characteristic morphology (Figure 1B). Adherent ATDCs showed round morphology, prominent cytoplasm, and short extensions and clustered together. Suppressor macrophages were adherent but did not cluster, showed a central body and some long extensions. A heterogeneous population of MDSC was obtained, comprising cells of various sizes, both adherent and nonadherent, some clustering together.

Furthermore, the yield of recovery varied between each RMC type. Indeed, whereas it reached 70% and 50% for MDSCs and ATDCs, respectively, suppMφ represented only 20% of the initial cell number (Figure 1C).

In Vitro–Derived RMCs Display Different Phenotypes

To define each RMC type, we analyzed the expression of surface markers. All 3 RMC types expressed CD11b, a myeloid cell marker (Figure 1D). Autologous TolDC and suppMφ also expressed CD11c and low/intermediate level of MHC class II, whereas MDSCs were negative for these 2 markers. As previously described,^27,28 Myeloid-derived suppressor cells expressed Gr-1 and were negative for F4/80. Suppressor macrophages displayed a typical macrophage phenotype, expressing F4/80, CD169, and CD64, whereas Ly6C and Ly6G were absent. Lastly, as previously described, ATDCs were CD11c^high, MHC class II^low, shared some markers with monocyte-derived inflammatory DCs, including FcεRI and F4/80^10,38 and did not express CD169 (Figure 1D and Table 1). Consistent with an immature state, all 3 types of RMC expressed low levels of the costimulatory molecule CD86, in contrast to control myeloid cells (Figure 1D).

TABLE 1: List of markers assessed on RMCs

RMC Subsets Display Different Intrinsic Properties

Regulatory myeloid cells were exposed to LPS to assess resistance to maturation, an important characteristic of RMCs. As shown in Figure 2A, all RMCs were resistant to maturation because they did not significantly upregulate MHC class II, the activation marker, CD40, or the costimulatory markers CD80 and CD86.

FIGURE 2: Maturation resistance and antigen endocytosis/degradation properties of the different RMCs. A, Expression of maturation-associated markers (MHC class II, CD80, CD86, and CD40) by RMCs under normal conditions (medium) and after 48 hours of LPS stimulation (n = 3). B, Representative histogram (left) and kinetics (right) of RMC antigen endocytosis (OVA-AF647) and degradation (DQOVA). Data are representative of 3 independent experiments.

Another intrinsic property of myeloid cells is antigen endocytosis and degradation, which precede presentation to T cells. Using OVA beads, we determined that ATDCs and suppMφ displayed high endocytic capacity which increased over time (Figure 2B). Strikingly, only a small fraction of MDSCs could endocytose antigens, even 24 hours after antigen exposure. To evaluate the degradative capacity of RMCs, cells were incubated with DQ-OVA, which emits fluorescence only after degradation. Autologous TolDCs and suppMφ were able to efficiently degrade antigens (Figure 2B). Degradation was also detected in the small population of endocytic MDSCs.

Only ATDCs and SuppMφ Induce T Cell Hypoproliferation, But All 3 RMC Types Are Suppressive

To examine the mechanisms of immunoregulation of each RMC type, we analyzed the ability of C57BL/6-derived RMCs to stimulate CFSE-labeled Balb/c T cell proliferation. As shown in Figure 3A, ATDCs and suppMφ did not induce CD4⁺ or CD8⁺ T cell proliferation at any RMC:T cell ratio. Although MDSCs did not induce T cell proliferation at high RMC:T cell ratios, T cell proliferation was observed at lower ratios (Figure 3A). Low levels of IFN-γ were detected in MDSC coculture supernatant, whereas it was not detectable in ATDC and suppMφ cocultures (Figure 3B). IL-10 secretion was not detected in any coculture supernatant (data not shown).

FIGURE 3: Suppression of T cell proliferation by the different RMCs. A, CFSE-labeled purified T cells from naïve Balb/c mice were cultured with C57BL/6 RMC or control cells for 4 days. Proliferation of CD3⁺CD4⁺ or CD3⁺CD8⁺ T cells in allogeneic co-culture was assessed by CFSE dilution using flow cytometry. Data are representative of 3 independent experiments. B, IFN-γ production was measured in the culture supernatant at day 4 (1:8 ratio) (n = 4 experiments). C, CFSE-labeled purified T cells were cultured with anti-CD3/28 beads in the absence or presence of syngeneic RMCs (ratio 2:1) for 4 days. Representative histograms show CD3⁺CD4⁺or CD3⁺CD8⁺ T cell proliferation. D, A bar chart depicts the mean proliferation ± SEM. Statistics were calculated by the nonparametric Mann-Whitney U test (*P < 0.05, ***P < 0.0001, ND, not detectable) n=4 mice per group from 4 independent experiments. E, IFN-γ production was measured in the culture supernatant at day 4 (n = 2 experiments).

To evaluate the suppressive capacity of RMCs, CFSE-labeled C57BL/6 T cells were stimulated with αCD3/CD28-coated microbeads in the presence of syngeneic RMCs. All RMCs were able to inhibit syngeneic polyclonal T cell proliferation, but ATDCs and suppMφ appeared to be more potent than MDSCs in inhibiting both CD4⁺ and CD8⁺ T cells (Figures 3C and D). Furthermore, no IFN-γ or IL-10 was detected in the supernatant of RMC cocultures (Figure 3E and data not shown).

RMC Subsets Use Distinct Mechanisms of T Cell Suppression

To further investigate RMC mechanisms of action, their effects on T cells were analyzed after allogeneic cocultures. As shown in Figure 4A, ATDCs did not induce the expansion of CD4⁺Foxp3⁺ Treg cells, whereas both suppMφ and MDSCs were able to increase the percentage of Treg cells. As previously described,^39-42 control myeloid cells were also able to generate or expand CD4⁺Foxp3⁺ Treg cells (Figure 4A), consistent with their semimature phenotype.

FIGURE 4: Suppression of T cell proliferation by RMCs is mediated by different mechanisms. Purified Balb/c T cells were cultured with C57BL/6 RMC or control cells for 4 days (ratio RMC:T cells 1:8). Graphs depict the mean ± SEM of (A) CD4⁺Foxp3⁺ Treg cells in CD3⁺ cells, (B) DAPI⁺ dead cells in CD3⁺CD4⁺ (left) and CD3⁺CD8⁺ T cells (right), and (C) CD69 (top) and CD25 (bottom) activation markers in CD3⁺CD4⁺ (left) or CD3⁺CD8⁺ (right) T cells as determined by flow cytometry. Statistics were calculated using the nonparametric Mann-Whitney U test (*P < 0.05, **P < 0.001) with n = 3-4 mice per group from 3 independent experiments.

Induction of apoptosis is another key mechanism by which T cell proliferation is regulated. Myeloid-derived suppressor cells induced a significant increase in cell death in both CD4⁺ (left panel) and CD8⁺ (right panel) T cells, compared with other cell types (Figure 4B). Low expression of CD25 and CD69 together with the lack of proliferation (Figure 3A) are indicative of minimal T cell activation in ATDC cocultures, in contrast to the other RMC types (Figure 4C).

To determine whether ATDCs were responsible for the unresponsiveness of allogeneic T cells, a 2-step stimulation experiment was performed. Allogeneic T cells were purified after coculture with each RMC type and restimulated with mature allogeneic splenic DCs (Figure 5A). CD4⁺ and CD8⁺ T cells cultured with suppMφ and MDSCs were able to proliferate in response to a second allogeneic stimulation, whereas ATDC-cultured T cells displayed decreased or no proliferation (Figure 5B). Moreover, a decrease in IFN-γ production was observed in the supernatant of ATDC-cultured T cells compared with other cocultures (Figure 5C). No IL-10 secretion was detected in any condition (data not shown).

FIGURE 5: ATDCs induce T cell unresponsiveness. CFSE-labeled purified Balb/c T cells were cultured with C57BL/6 RMC or control cells (ratio 1:1) for 4 days. T cells were then isolated using magnetic beads and restimulated with irradiated C57BL/6 splenic mature DCs for 3 additional days (ratio 4:1). A, Outline of the experimental design. B, Proliferation of CD3⁺CD4⁺ or CD3⁺CD8⁺ T cells was assessed by CFSE dilution using flow cytometry. C, IFN-γ secretion at the end of the second culture was assessed in supernatants by ELISA. Statistics were calculated by the nonparametric Mann-Whitney U test (*P < 0.05, **P < 0.001) n = 3 mice per group from 3 independent experiments.

Together, these results demonstrate that ATDCs induce T cell hyporesponsiveness, whereas suppMφ and MDSCs induce Treg cell expansion. In addition, MDSCs induce T cell death.

Adoptive Transfer of Autologous RMCs Prolongs Graft Survival

To assess the in vivo efficacy of each type of RMC in an autologous setting, we used a mouse model in which male C57BL/6 skin was grafted onto female recipients. Regulatory myeloid cells were intravenously injected the day before transplantation. It is important to note that recipients were fully immunocompetent, and no immunosuppressive treatment was administered. Injection of 1 million ATDCs significantly prolonged graft survival (median, 31 days), whereas an equivalent number of suppMφ did not prolong graft survival compared to untreated mice (median, 25 days vs median 23.5 days in untreated mice) (Figure 6A, left panel). Surprisingly, injection of 3 million ATDCs did not prolong allograft survival (median, 25 days) (Figure 6A, right panel). This difference may be explained by some unexpected early rejections (3 mice among 12 mice) that occurred in this group. However, when 3 million suppMφ were injected, graft survival was prolonged (median, 27 days). A single injection of autologous MDSCs prolonged graft survival when 1 million cells were injected (median, 28 days), and this effect was enhanced by injection of 3 million cells (median, 45 days) (Figure 6A).

FIGURE 6: Adoptive transfers of ATDCs, suppMφ or MDSCs prolong graft survival. (A) Curves representing skin graft survival of n = 5-16 mice per group pooled from > 3 independent experiments. Survival rates were analyzed by Log-rank (Mantel-Cox) test (*P < 0.05, **P < 0.01). Interquartile range (days): untreated mice, 4.5; 1 million ATDC, 10.75; 1 million suppMφ, 10.75; 1 million MDSC, 12; 3 million ATDC, 10.5; 3 million suppMφ, 18; 3 million MDSC, 23.5. B, C57BL/6 female mice received one million (ATDC) or three million (suppMφ or MDSC) autologous RMCs 1 day before the transplantation of C57BL/6 male skin. Untreated mice were grafted but did not receive RMCs. dLN (left) and skin grafts (right) were recovered 14 days after transplantation, and the percentages of CD8⁺ UTYpentamer⁺ cells present in CD3⁺CD8⁺ cells were determined by FACS staining. Percentages analyzed by the non-parametric Mann-Whitney U test (**P < 0.01). n = 5 mice per group.

To study the mechanisms involved in the improved graft survival following RMC treatment, we performed adoptive cell transfer of 1 million ATDCs and 3 million suppMφ or MDSCs. Recipient mice were sacrificed 14 days after transplantation, and immune cells derived from spleen, draining lymph nodes (dLN), and skin grafts were analyzed. Transplanted mice showed increased absolute cell numbers in the dLN (but not in the spleen) compared with naive nontransplanted mice. However, no significant differences were observed between the RMC treatments in spleen, dLNs, or skin (Figure S1A, SDC, https://links.lww.com/TP/B305). Therefore, RMCs do not seem to prevent leukocyte recruitment or proliferation in the dLNs or within transplanted tissue at 14 days posttransplantation.

A more in-depth analysis of immune cell populations performed 14 days posttransplantation did not highlight any differences in the percentages of CD3⁺, CD19⁺, CD4⁺, or CD8⁺ cell populations in the spleen, dLNs, or skin graft (Figure S1B, SDC, https://links.lww.com/TP/B305). At the same time point, no difference in the percentages of CD4⁺Foxp3⁺ Treg cells or dead T cells was observed (data not shown). Furthermore, expression of the activation markers, CD25 and CD69, was similar among all mice (data not shown). However, we detected an increased percentage of antigen-specific CD8⁺ T cells in dLN, specifically in mice treated with ATDCs (Figure 6B, left panel) as well as mice with MDSC-treated skin grafts (Figure 6B, right panel), the latter representing up to half of the total CD8⁺ T cell population. We speculate that this population are regulatory CD8⁺ T cells, as previously reported.¹⁰ In addition, our data indicate that ATDCs and MDSCs may induce antigen-specific CD8⁺ T cells, with distinct kinetics and patterns of migration.

DISCUSSION

In this study, we aimed to characterize 3 in vitro-derived RMC types. Our results highlight the unique phenotype and functionality of each subset. Autologous TolDCs and suppMφ express low levels of costimulatory molecules, correlating with their capacity to inhibit T cell proliferation and IFNγ secretion. Interestingly, MDSCs are CD86^lowCD80^high and are associated with an intermediate induction of T cell proliferation and IFNγ secretion. The 3 RMC subsets exert their effects on T cells in vitro through different mechanisms; ATDCs control activation, proliferation, and reactivation of T cells, whereas suppMφ induce a state of nonproliferative “altered” activation associated with Treg cell expansion/induction. Lastly, MDSCs appear to exert their immunosuppressive function through Treg cell expansion as well as induction of T cell death, potentially caused by overactivation. Furthermore, our results demonstrate for the first time that autologous suppMφ are able to promote graft survival, whereas the potency of autologous ATDCs and MDSCs to prolong graft survival has already been reported by our group and others.^10,27,28

Autologous TolDCs were differentiated following a protocol described by Lutz et al⁵ and we confirmed that GM-CSF TolDCs induce T cell hyporesponsiveness and anergy.^5,43 Interestingly, the authors reported that induced anergic T cells could be converted to Tr1-like Treg cells after a second stimulation with immature tolerogenic DCs.⁴³ In our hands, neither IL-10 nor Foxp3⁺ Treg cell expansion was detected in vitro, whereas other studies on TolDCs have reported their capacity to either promote the expansion of natural CD4⁺CD25⁺FoxP3⁺ Treg cells^19,44 or induce Treg cells from naive CD4⁺CD25⁻ T cells.⁴⁵ However, in accordance with our previous study, we confirmed that injection of ATDCs promotes skin graft survival in association with an expansion of antigen-specific regulatory CD8⁺ T cells in dLNs.¹⁰ We previously reported in a model of islet transplantation that the combination of ATDC injection and short-term anti-CD3 treatment induces allograft tolerance associated with an expansion of Foxp3⁺ Treg cells in lymphoid organs.⁶ Other tolerogenic DCs, such as RAPA-DCs or RelB-silenced DCs, are also known to promote indefinite prolongation of cardiac allograft survival associated to Foxp3⁺ Treg cell expansion.^18,46

Myeloid-derived suppressor cells generated in this study comprised a heterogeneous population containing CD11b⁺ Gr1⁺ and CD11b⁺Gr1^low cells in accordance with the initial report by Marigo et al.²⁷ Compared with ATDCs and suppMφ, MDSCs exhibit a reduced suppressive capacity and partially stimulate allogeneic T cells. Further experiments will be necessary to elucidate whether some MDSC subsets display greater suppressive potential than others. For example, CD11b⁺Gr1⁺ cells were previously described as immunogenic and could dampen the suppressive activity of the whole MDSC population.⁴⁷ Myeloid-derived suppressor cells obtained by other differentiation protocols, for instance, including the combination of GM-CSF and IL-13,²⁵ were also reported to inhibit T cell proliferation. Regulatory T cell expansion induced by MDSCs is well documented in cancer.⁴⁸ In this line, we showed here that in vitro culture of MDSCs with T cells promoted the expansion of Foxp3⁺ Treg cells. These results were also reported in vivo where injection of GM-CSF and hepatic stellate cell differentiated MDSCs promoted islet allograft survival in association with an induction of Foxp3⁺ Treg cells.⁴⁹ Here, injection of autologous GM-CSF+IL-6-differentiated MDSCs was also able to prolong skin graft survival following a single injection of cells.

Lastly, suppMφ display high suppressive capacity and favor T cell hyporesponsiveness. Few reports on in vitro–differentiated regulatory/suppressor macrophages are available in the literature. Regulatory macrophage cells, characterized by Riquelme et al,¹¹ were differentiated from BM-purified monocytes with M-CSF, human serum, and brief activation with IFN-γ. The authors showed that IFN-γ–stimulated Mreg display in vitro regulatory activity and prolong cardiac allograft survival. Suppressor macrophages used here were derived without activation/maturation stimuli, similar to ATDC and MDSC protocols. Because recent literature supports a role for maturation/activation of macrophages in their regulatory ability,^11,29 we could hypothesize that suppMφ are precursors of Mreg and require activation to exert their function. Even if no significant maturation of suppMφ was observed after LPS stimulation, the trend toward upregulation of MHC Class II and CD40 molecules observed here (Figure 2A) suggests a potential activation of these suppMφ. Given that recipient derived suppMφ prolong skin graft survival, it would therefore be interesting to stimulate suppMφ with IFN-γ before injection to evaluate their efficacy under these conditions. Riquelme et al¹¹ demonstrated the potency of donor Mreg, injected 8 days before transplantation, in prolonging graft survival whereas recipient Mreg do not induce any graft protection. Growing evidence show that the timing of cell administration is a key aspect in the context of cell therapy in transplantation. Indeed, our previous experiments suggest that autologous ATDCs firstly migrate to the graft, capture antigens, and then home to dLNs to present antigens to T cells.¹⁰ The time of injection of cells is therefore crucial, because early injection may not culminate in cell migration to the appropriate site. Autologous cell therapy presents safe and practical advantages over the use of donor-derived RMCs.³¹ Furthermore, it was recently reported that endogenous DC-SIGN⁺ suppressor macrophages accumulating in the allograft were able to promote the induction of transplant tolerance following α-CD40L monoclonal antibody administration in a fully mismatched model of cardiac transplantation.⁵⁰ Notably, these suppressor macrophages display some common features with suppMφ, such as developmental dependence on M-CSF, expression of the surface marker, CD209, and the ability to promote Foxp3⁺ Treg cell expansion. Further experiments will be necessary to determine whether suppMφ could represent in vitro counterparts of DC-SIGN⁺ suppressor macrophages.

In summary, our data demonstrate that 3 populations of RMCs can be distinguished by their morphology, phenotype, intrinsic properties, and in vitro mechanisms of action on T cells. Further studies will be necessary to elucidate the mechanisms by which these cells exert their functions in vivo.

ACKNOWLEDGMENTS

The authors kindly thank Timothy Murray for editorial revision of the article.